IEEE Trans Neural Syst Rehabil Eng. 2019 May;27(5):876-886. doi: 10.1109/TNSRE.2019.2908817. Epub 2019 Apr 2.
Although recent advances in neuroprostheses offer opportunities for improved and intuitive control of advanced motorized and sensorized robotic arms, practical complications associated with such hardware can impede the research necessary for clinical translation. These hurdles potentially can be reduced with virtual reality environments (VREs) with embedded physics engines using virtual models of physical robotic hands. These software suites offer several advantages over physical prototypes, including high repeatability, reduced human error, elimination of many secondary sensory cues, and others. There are limited demonstrations of closed-loop prostheses in the VRE, and it is unclear whether VRE performance translates to the physical world. Here we describe how two trans-radial amputees with neural and intramuscular implants identified objects and performed activities of daily living with closed-loop control of prostheses in the VRE. Our initial evidence further suggests that capabilities with virtual prostheses may be predictors of physical prosthesis performance, demonstrating the utility of VREs for neuroprosthetic research.
尽管神经假体的最新进展为改进和直观控制先进的机动和感应机器人手臂提供了机会，但与这种硬件相关的实际并发症可能会阻碍临床转化所需的研究。这些障碍可以通过使用物理机器人手的虚拟模型的虚拟现实环境 (VRE) 与嵌入式物理引擎来降低。这些软件套件与物理原型相比具有几个优势，包括高重复性、减少人为错误、消除许多次要感觉提示等。在 VRE 中对闭环假肢的演示有限，并且不清楚 VRE 的性能是否转化为物理世界。在这里，我们描述了两名具有神经和肌内植入物的桡骨截断患者如何在 VRE 中使用闭环控制假肢来识别物体和进行日常生活活动。我们的初步证据进一步表明，虚拟假肢的能力可能是物理假肢性能的预测因素，这证明了 VRE 对于神经假肢研究的实用性。
